In his arguments, Donellan proposed a semantic account (though he denied it) to differentiate between the attributive and referential uses of language. This differentiation was illustrated through the "martini glass" example: When a partygoer asks "Who is the person drinking the martini?", he is using the act of drinking the martini as a place holder to refer to the man. Even if the man was not drinking a martini, the meaning of his sentence would stay the same. However, if the chairman of the Teetotalers Union uttered the same statement, the act of drinking of martini is all important. If it turned out the man was in fact drinking water, the chairman would be referring to nobody. Thus, he used the same sentence to attribute. In the attributive use, the characteristic is of significance, while in the referential use it is merely a place holder. If all the conditions set forth in a statement fail to be met, the attributive use would have a zero truth value, while the referential case may still be true.
Through the referential/attributive argument, Donellan claims to refute Russell and Strawson with an argument that is not semantically ambiguous. However his aforementioned argument is based on semantic ambiguity (the speaker uses two sentences with the same semantic content to express different ideas.) Kripke later 'corrects' Donellan's theory by distinguishing between pragmatics and semantics.
Kripke provides a pragmatic account to explain the attributive/ referential distinction found by Donellan. By using a pragmatic account, Kripke was able to justify elements of both Russell's and Donellan's theories. First, Kripke pointed out that in the logically precise Russell languages, semantic ambiguity was eliminated. Therefore, any distinctions found were the result of something other than semantics, and Donnellan's attributive/referential distinction should be eliminated. However, Kripke did observe a case of ambiguity: arch uses. In these uses, a general term is used to refer to a specific being. When analyzed using semantics, it is impossible to discern what is being referred to, and the Russell language interpretation remains the same for two different meanings. Since the Donellan distinction was still observed in non-semantically ambiguous sentences, the attributive/referential distinction must lie outside the realm of semantics. Thus Kripke introduced the idea of speech acts providing the pragmatic difference. Depending on the situation and manner of speech, a sentence may have various meanings. As an example, Kripke noted that a burglar's utterance of "The cops are around the corner" means "Let's split". Thus, the actual meaning of the words in a given context is much different than the direct semantic interpretation of the same words.
Kripke concluded that Donnellan and Russell could be justified with each other. However, Russell was long since dead, and was more worried about other things in his life. Frege and Russell attempted to construe singular sentences to express general propositions. Thus, the quantified language of Russell had most sentences reduced to a series of predicates expressing general propositions. Putnam and Kaplan borrowed from points of the project to construct their own theories that didn't rely on internal meanings and general propositions.
Putnam tried to show that meaning is not in the head, through the use of various examples. One of his more interesting examples was the "Twin Earth" example. Here, he described a planet identical to earth, with the only difference being that water is represented by the complex compound XYZ instead of H2O. (Unfortunately, this would result in a vastly different universe, for many chemistry problems would be different due to the different makeups and atomic weights. Also, the forces present in the makeup of XYZ may also cause additional [but yet undiscovered] bonding characteristics that affect twin-earth in isolated ways. However, for the sake of his argument we can assume the difference is minor to the extent that the atomic weight remains the same, and the bonding characteristics do not differ in any noticeable ways from what we know as water.) When a person from earth sees XYZ on Twin Earth, he may mistakenly refer to it as water. However, after analysis, he will realize that it is not what is commonly referred to as 'water' on earth, but instead is the substance XYZ. But what if the people lived in the days before we were able to identify the chemical makeup of compounds? In that case, they would still be referring to different things, but their corresponding psychological states would be identical. Thus Putnam concludes that meaning "is not a function of the psychological state of the speaker by itself." Instead meaning relies on the knowledge of experts. The experts identify a certain object as water. Later, when referring to water, an individual is merely referring to the same thing that the experts typed. The experts identity takes precedent over the individual.
Putnam used the 'beech-elm' case to further his hypothesis. Someone may be unable to distinguish an elm tree from a beech tree. Thus, when he mentions "elm" he may be referring to either a beech or an elm tree. His psychological state will be the same whether he is referring to one or the other. However, another person may be able to distinguish between the two, and attempt to correct the first person. Unfortunately, the first person, has no concept of the difference between the two, and will still erroneously refer to beeches as elms. Even though beeches are beeches and elms are elms, the first person refers to them as the same. Even though the data in his head remains the same, the objects are in fact difference. Therefore, meaning is not in the head, but dependent upon the universal view.
Putnam explanations are explained using the "linguistic division of labor" and "quasi-indexical character of meaning" concepts. With the linguistic division of labor, an expert decides upon the original meaning, while other people merely use what is necessary for their identification. For example, a chemist may have knowledge of many detailed aspects of gold, such as melting point, electronegativity, enthalpy, etc. that he used to exactly identify gold. A common man, however, just relies on a few basic observations to single out the gold. Thus the linguistic labor was divided. Furthermore, the common man would use gold conclusively identified by the expert as the 'indexical' to base his definition of gold.
Kaplan proceeded in the same fashion to attempt to give a Fregeian account of demonstratives and indexicals. With demonstrative he could give a Fregeian account using the actual object being pointed to in the sense. For example "That Cow is eating grass" would contain the actual act of pointing to the cow in the meaning of the sentence. However, this did pose a problem with indexicals, where no act of pointing existed. Thus indexicals could contain a 'virtual' pointing, in which the general object is to be understood. Using this method, the interpretation of "That's Fred" and "I am Fred" would be essentially the same, with the pointing unnecessary in the indexical ("I am Fred") case due to the general understanding of the audience. Unfortunately, Frege's puzzle still exists. For example, if Bill's utterance of "someone's making a mess" has a different meaning once corrected to "I am making a mess". Each sentence could be interpreted as "Bill is making a mess", but the second one is the result of further study, and has a different meaning. Thus Kaplan has some qualms about the Fregeian 'sense' account.
Kaplan also gave a direct reference account of reference. Instead of having the act of pointing in a proposition, the actual object was in the proposition. Kaplan liked his demonstrative-indexical count so much that he also decided it would be a great way of explaining names. However, using this theory, meaning would be based on existence. If something didn't exist, its proposition would contain a null character. However, there still remains the problem of how sentences such as "Does Santa Clause exist?" are to be interpreted. If Santa Clause did exist, the sentence would have a very different meaning than if he didn't. However, the sentence "Is Santa Clause the Easter Bunny" would have both beings translated to null, and thus conclude that Santa was rabbit, which seems very unlikely. (But it does provide a great chance for future research project...)
- [For Philosophy of Lanaguage] 4/18/93